Dapr’s microservices runtime now supports AI agents

Back in Microsoft in 2019 open source daperMicroservis based applications are easier for a new business time to build. At the same time, no one talked about AI agents, but when it came out, there were some of the fundamental building blocks to support Early Mounted AI agents. Because one of the main features of DAP is a virtual concept the actorsMessages that can receive and process messages regardless of all other actors in the system.

Today, DAPR started DAPR agents, which helps to build AI agents by providing many construction blocks to do so.

“Agents are a very good use for DAPR,” said DPR Co-Creator and Mainerer Yarn Schneider. “From a technical point of view, you can use the actors very light to use it and be able to manage them on a scale of the state. And many people are just a matter of work, but then you need to write a lot of work. The same level of orchestra and DAPR is not killed. “

DAPR agents are origin SlidingThis is a popular open source project that expands DAP for the AI ​​Agent. Project protectors, including Microsoft AI researcher Roberto Rodriguez, including two teams decided to bring the project under the DPR umbrella to ensure the sustainability of the new agent.

“Many ways of agent systems and all terminology around them are” distributed systems’, together for ‘distributed systems’, the joint creative and protective Mark Fussell said. “[…] Instead of calling them mikloservices, you can now call them agents so you can put great language models inside large language models. “

You need an orchestra engine and benefit to coordinate agents effectively, the team argues – this is the fullest of DAPR. Because when the actors of DAP are extremely effective and a message in the milliseconds (and rotated in the form of protection, protected by their closure, and scroll in MiliceNeber.

Currently, DAPR agents can talk outside the box for most popular model providers. These include AWS Bedrock, Openai, Anthropic, Mistral and hugs. Support for local LLMS will come very soon.

On communication with these models, DAPR agents have the opportunity to determine the list of tools that develops the existing DAPR framework, developers, the agent can use the agent.

Currently, DAPR agents will soon began to support Python with .Net support. Java, JavaScript and Go will follow soon.